CLAP4 CLIP: Continual Learning with Probabilistic Finetuning for Vision-Language Models

This makes them overlook the many possible interactions across the input modalities and deems them unsafe for high-risk tasks requiring reliable uncertainty estimation.